Putting pretrained models to work on images, speech and time series.
Trend, seasonality, honest baselines and time-based testing, using NASA temperatures and bike rentals from the hub.
Transcribe and translate audio with OpenAI's open Whisper model, handle long recordings, and measure accuracy with word error rate.
Classify images with a pretrained ResNet-50 in ONNX Runtime, then adapt a pretrained network to your own categories.
35 in this topic
Classifying images without training from scratch: choosing a pretrained model, preparing images correctly and reading the output.
Applied AI 2 min read 19 May 2026
How object detection finds and localises multiple objects in an image, the main model families and how accuracy is measured.
Applied AI 2 min read 18 May 2026
Semantic, instance and panoptic segmentation: labelling images pixel by pixel, and where each is used.
Applied AI 2 min read 17 May 2026
Extracting text from images and scanned documents: how modern OCR works, where it fails and how to improve results.
Applied AI 2 min read 16 May 2026
Finding people, organisations, places and other entities in text: approaches from rules to transformers and LLMs.
Applied AI 2 min read 15 May 2026
Classifying opinion in text as positive, negative or neutral, the main approaches, and the pitfalls of sarcasm, domain and aggregation.
Applied AI 2 min read 14 May 2026
From keyword rules to fine-tuned transformers and LLMs: how to build a classifier for tickets, emails or documents.
Applied AI 2 min read 13 May 2026
How AI translation works today, its strengths and weaknesses, and how to use it safely for business content.
Applied AI 2 min read 12 May 2026
Extractive, generative and retrieval-based question answering: how each works and how to choose.
Applied AI 2 min read 11 May 2026
Turning invoices, forms and contracts into structured data with OCR, layout models and LLMs — and keeping humans in the loop.
Applied AI 2 min read 10 May 2026
Forecasting sales and demand: data you need, methods from baselines to machine learning, and how to measure and use forecasts.
Applied AI 2 min read 9 May 2026
How machine learning detects fraud in payments and claims, handling rare events and adversaries, and balancing losses against friction.
Applied AI 2 min read 8 May 2026
Predicting which customers are likely to leave, defining churn properly, and turning predictions into effective retention actions.
Applied AI 2 min read 7 May 2026
Using sensor data and machine learning to predict equipment failures before they happen — data, approaches and pitfalls.
Applied AI 1 min read 6 May 2026
Where AI is helping in healthcare — imaging, documentation, operations — and why validation, bias and regulation demand care.
Applied AI 2 min read 5 May 2026
How machine learning and language models help security teams detect threats and respond faster — and how attackers use AI too.
Applied AI 1 min read 4 May 2026
Building personalised experiences that users value: data, cold start, evaluation with experiments, and avoiding creepiness.
Applied AI 1 min read 3 May 2026
Using cameras and AI to spot defects on production lines: setup, data collection, model choices and running it reliably.
Applied AI 1 min read 2 May 2026
Analysing maps and satellite images with machine learning: data sources, common tasks and pitfalls with location data.
Applied AI 2 min read 1 May 2026
How learning-to-rank improves search results, the signals it uses and how to evaluate ranking changes safely.
Applied AI 1 min read 30 Apr 2026